Maid
An open source Android AI chat app that runs GGUF models offline through llama.cpp and connects to remote providers, including Ollama and OpenAI.
Maid is an Android AI chat app for people who want to run models on their phone and access remote models in the same app. It runs GGUF models locally through llama.cpp without an internet connection. It's open source under the MIT license, with no ads or telemetry.
You can choose models from a curated Hugging Face catalogue, including Qwen, Phi, LFM and TinyLlama, or load a GGUF file you already have. Local inference keeps model processing on the device. Remote conversations go to the provider you connect, with support for Anthropic, DeepSeek, Mistral and OpenAI, as well as Ollama and llmman. Maid uses your own API key for remote providers.
The chat controls go beyond choosing a model. You can adjust generation settings and context length for each session, and set a global system prompt to define the assistant's behaviour. Maid lets you organise conversations and export or import them as JSON, so chat history can move outside the app.
Account sync is optional. If you enable it, Supabase backs up your settings and chat history, adding a cloud component to an otherwise on-device workflow. Voice output comes through the companion app Maise. The interface uses Material You styling, with light and dark themes that follow your system preference.
